Hello everyone,   
      
   I have a large database of classified ads. Each contain a text (Less   
   than 1000 chars) that describes the product on sale, and a price.   
      
   E.g.   
      
    - Bicycle, red, almost new, 26" frame. : Price: 2500 Categoy : Bicycle   
      
    - Car, Ford, 1.6, 1999, Red metallic : Price: 45000 Category : Cars,   
   Ford   
      
    - Car, Ford, 1.6, 1999, Blue metallic : Price: 47000 Category : Cars,   
   Ford   
      
      
   The Category is defined in the database as a descrete variable.   
   One will never find a bicycle in the "Cars" section.   
      
   I am doing som resarch into the possibility of using text mining to   
   quantify each word as a vector influincing the price. In the example   
   (wich is quite realistic) one can see that Blue metallic will influince   
   the price positively, in relation to Red Metallic.   
      
   The Category field will be used as a dimension. Following the example   
   the word "red" will have less or none influince in a bicycle, but more   
   influince on a car.   
      
   The text categorization will have to be independent of language, so the   
   basic term extraction of SSIS will not cut it. In fact all words in the   
   ad might have some importance.   
      
   What methodology will I use ? I'm sure this can be implemented.   
      
   Will you take the challange and help me get closer to a solution ?   
      
   kind regards   
      
   Mads   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)